The massive diffusion of smartphones we are seeing in the last years, the growing interest for everything related to wearable devices and Internet of Things (IoT), the exponential rise of Location Based Services (LBS: services based on the position of the users and accessible through the device) has meant that technologies capable of determining the position of the user inside a specific context have taken a crucial role in the consumer sector. In outdoor environments, GPS (Global Positioning System) can be considered today as a standard de facto , while the localization and navigation in indoor environments still remains one of the technological challenges of the next years. Indoor Positioning Systems (IPS) have a remarkable importance in a lot of important market segments such as the retail sectors for con- textual advertising (commercial centers, supermarkets), touristic and transportation sectors (airports, museums), healthcare sectors (hospitals). Sometimes, in emergency situations they can make the difference between life and death. Even if actually doesn t exist a definitive solution as efficient and precise as GPS and with all its advantages, various approaches and methodologies has been proposed in the last years in scientific literature, and several technologies are appearing into the market. The researches have focused, rather than on realizing a general purpose IPS with high performances everywhere, on the development of a variety of solutions suitable for the specific place of deployment and which meet the specifics precision, security, invasiveness and cost requirements. In the first part of this doctoral dissertation, after an overview on the main methodologies used for locating a user in an indoor environment, we will analyze the state of the art and present some of the most interesting cases. We will focus particularly on approaches which utilize the smartphone s inertial sensors and 2D visual markers based computer vision techniques. In the second part of the dissertation we will propose our own solution to the indoor localization problem. Such solution consists in a visual markers system deployed onto the area of interest s floor and a step detection algorithm. We will compare the performances of three types of markers, taken from the scientific literature, from the point of view of the specifications required by the suggested solution, and choose the most better performing one. We will then propose a client - server architecture for managing the whole process of tracking the user inside a building, and present an implementation of the client, on the iOS platform. Finally in the last part of the dissertation, we will show the obtained results, and talk about the possible future works.
L'elevata diffusione degli smartphones alla quale stiamo assistendo negli ultimi anni, il crescente interesse per tutto ciò che riguarda i wearable devices e l internet of things, l’aumento esponenziale di servizi basati sulla posizione (Location based services - LBS) e fruibili attraverso il terminale, ha fatto sì che le tecnologie capaci di determinare la posizione dell’utente all’interno di uno specifico contesto abbiano assunto un ruolo di primaria importanza nel panorama tecnologico consumer attuale. Per quel che riguarda l’ambiente outdoor, il GPS può essere considerato uno standard de facto, mentre la localizzazione e la navigazione in ambiente indoor rimane una delle più importanti sfide tecnologiche dei prossimi anni. Gli indoor positioning systems (IPS) hanno un’importanza notevole in tutta una serie di importanti segmenti di mercato quali il settore retail (centri commerciali, supermercati) per il contextual advertising, il settore turistico (musei, aeroporti), il settore medico-assistenziale (ospedali, centri di assistenza anziani). A volte, in situazioni di emergenza, gli IPS possono fare la differenza tra la vita e la morte. Nonostante attualmente non sia disponibile una soluzione definitiva che eguagli l efficienza e la precisione del GPS e ne conservi tutti i vantaggi, vari approcci e metodologie sono state proposte negli ultimi anni in letteratura scientifica, e varie tecnologie sono apparse sul mercato. Ci si è concentrati, piuttosto che sul realizzare un IPS general purpose che andasse bene dappertutto, sulla messa a punto di molteplici soluzioni che si adattassero al particolare luogo di deploy ed alle specifiche di precisione, sicurezza, invasività e costi. Nella prima parte della presente tesi di dottorato, dopo una overview delle principali metodologie utilizzate per la localizzazione di un utente in ambienti indoor verrà analizzato lo stato dell’arte e presentati alcuni casi di maggior interesse. Ci si concentrerà in particolare su approcci che utilizzano i sensori inerziali presenti all’interno degli smartphone e tecniche di computer vision basate su marker 2D visuali. Nella seconda parte della tesi verrà proposta una soluzione al problema dell’indoor localization. Tale soluzione prevede un sistema di markers 2D installati sul pavimento dell’area di interesse ed un algoritmo di step detection. Verranno confrontate le prestazioni di tre tipologie di marker 2D presenti in letteratura, in relazione alle specifiche richieste dall’architettura proposta, e scelto il marker più performante. Verrà proposta quindi una architettura client server per gestire l’intero processo di tracking dell’utente all’interno di un edificio, e presentata una implementazione del client su piattaforma iOS. Nella parte finale della tesi infine verranno esposti i risultati ottenuti, e si parlerà dei possibili sviluppi futuri del lavoro svolto.
Methodologies and Technologies for Indoor Localization: a Computer Vision Based Approach / LA DELFA, GAETANO CARMELO. - (2015 Dec 09).
Methodologies and Technologies for Indoor Localization: a Computer Vision Based Approach
LA DELFA, GAETANO CARMELO
2015-12-09
Abstract
The massive diffusion of smartphones we are seeing in the last years, the growing interest for everything related to wearable devices and Internet of Things (IoT), the exponential rise of Location Based Services (LBS: services based on the position of the users and accessible through the device) has meant that technologies capable of determining the position of the user inside a specific context have taken a crucial role in the consumer sector. In outdoor environments, GPS (Global Positioning System) can be considered today as a standard de facto , while the localization and navigation in indoor environments still remains one of the technological challenges of the next years. Indoor Positioning Systems (IPS) have a remarkable importance in a lot of important market segments such as the retail sectors for con- textual advertising (commercial centers, supermarkets), touristic and transportation sectors (airports, museums), healthcare sectors (hospitals). Sometimes, in emergency situations they can make the difference between life and death. Even if actually doesn t exist a definitive solution as efficient and precise as GPS and with all its advantages, various approaches and methodologies has been proposed in the last years in scientific literature, and several technologies are appearing into the market. The researches have focused, rather than on realizing a general purpose IPS with high performances everywhere, on the development of a variety of solutions suitable for the specific place of deployment and which meet the specifics precision, security, invasiveness and cost requirements. In the first part of this doctoral dissertation, after an overview on the main methodologies used for locating a user in an indoor environment, we will analyze the state of the art and present some of the most interesting cases. We will focus particularly on approaches which utilize the smartphone s inertial sensors and 2D visual markers based computer vision techniques. In the second part of the dissertation we will propose our own solution to the indoor localization problem. Such solution consists in a visual markers system deployed onto the area of interest s floor and a step detection algorithm. We will compare the performances of three types of markers, taken from the scientific literature, from the point of view of the specifications required by the suggested solution, and choose the most better performing one. We will then propose a client - server architecture for managing the whole process of tracking the user inside a building, and present an implementation of the client, on the iOS platform. Finally in the last part of the dissertation, we will show the obtained results, and talk about the possible future works.File | Dimensione | Formato | |
---|---|---|---|
PhD Dissertation - Gaetano Carmelo La Delfa.pdf
accesso aperto
Tipologia:
Tesi di dottorato
Licenza:
PUBBLICO - Pubblico con Copyright
Dimensione
7.21 MB
Formato
Adobe PDF
|
7.21 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.